NetNews Usenet Archive 1992 #26/NN_1992_26.iso/spool/sci/math/stat/2325

Text File | 1992-11-08 | 1.7 KB | 35 lines

Newsgroups: sci.math.stat
Path: sparky!uunet!think.com!sdd.hp.com!news.cs.indiana.edu!noose.ecn.purdue.edu!mentor.cc.purdue.edu!hrubin
From: [e-mail] (Herman Rubin)
Subject: Re: Gallup Polls and Curving Grades
Message-ID: <[e-mail]>
Organization: Purdue University Statistics Department
References: <[e-mail]>
Date: Mon, 9 Nov 1992 02:03:59 GMT
Lines: 24

In article <[e-mail]> [e-mail] writes:
>Hello there,

>    I would like to get some information about gallup polls and  curving
>grades., i.e. the procedure that is followed to collect data and the thoery
>behind it. Do you have any suggested reference that I should refer to any book
>in sampling that does not get into too much technicality?

Any fair statistics book will give the distribution theory needed to get the
sampling error for a poll.  The other important thing is to get an accurate
sample, and the pollsters usually use a stratified random sample, with the
"random" being important.  Any fair text on sampling discusses this.

As for curving grades, this is only meaningful if one can assume that the
class is large enough and random enough that it reflects the population,
and that environmental factors, such as the teaching, the quality of class
discussion, etc., are irrelevant.  Even then, there is the question as to
why a particular distribution of grades should occur.  Grades mean little
enough without doing something with as little justification as curving.
--
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
Phone: [phone]
[e-mail] (Internet, bitnet)
{purdue,pur-ee}!snap.stat!hrubin(UUCP)